A motion-insensitive double-echo method to separate the effects of BOLD (blood oxygenation level dependent) contrast and inflow effects has been developed. Separation of the two effects is important for functional magnetic resonance imaging of brain activation (~RI) because although BOLD contrast is a short-range (tens of ~m) effect, the signal from inflow effects may be present far from the regions of activation. The mathematical decomposition of the effects requires that two images are acquired per excitation. From these images, Si and S2, two other images are computed, ~O and B. Only flow effects are present in the ~O image and only activation effects are present in the B image. Methods and Results Three in vitro experiments were conducted to demonstrate that BOLD and inflow effects could be separated using the dual-echo technique. The experiments were designed to produce IMRI-like signal changes by changing the flow rate in the phantom and by switching between fluids with the same Ti * but different T2* in the phantom. A flow phantom was constructed from a Plexiglas tube coupled to two fluid reservoirs using flexible plastic tubing. The flow rate was regulated by the height of the reservoirs and by restricting valves. In the experiments (1) flow rate was switched, (2) flow rate was held constnat and the fluid was switched between the two solutions, (3) flow rate and the fluid source were switched at different frequencies. Excellent separation of the BOLD and flow effects was achieved in all experiments. The flow effect was cleanly separated by the decomposition of the dual-echo spiral images. Discussion The in vitro experiments prove that the decomposition technique is a valuable tool for understanding the source of IMRI effects.